Present invention embodiments relate to information management, and more specifically, to processing group-by operations when some or all of the groups have a single member.
In data warehouse and operational analytics database queries, group-by operations (e.g., operations specified in a Structured Query Language (SQL) statement containing a GROUP BY clause) that result in a large number of groups are commonplace. Grouping may be performed by sorting input records on the grouping keys and collapsing the groups, or by maintaining a hash table that maps the grouping keys to aggregation records. These operations consume large amounts of processing and memory resources.